August 1, 2017
I have a confession to make. I sometimes ignore data. When my calorie-counting app tells me I only have 400 calories left for dinner, I have been known to eat the cheeseburger and the fries anyway. When my watch annoys me to get up from my desk and walk around, I usually keep typing.
It’s not that I specifically don’t care about my health. I ignore other things, too, like when my thermostat tells me it’s time to change the filter, or my operating system tells me there’s an update available, or my wife tells me to take out the trash. I realize I’m ignoring all of this data at my own peril, risking an inefficient HVAC system, outdated computer virus protection or an annoyed spouse. It’s just that, other than that last one, it’s easy to ignore data, especially when it’s not a priority.
The DIY Mindset
The challenges are multiplied for design engineers who have long had plenty of opportunities to ignore various design and simulation data points, and now software engineering data has entered the picture, and everyone’s talking about incorporating data coming in from connected devices into a product lifecycle management (PLM) workflow. Meanwhile, home-grown data management workflows that focus on high-priority data are still the norm for many engineers who either tried early product data management (PDM) software options and found them lacking, were able to cobble together their own data management software stack using what was already available to them, or were never able to convince the powers that be to invest in a purpose-built solution.
One issue with proprietary data management solutions is that they often rely on “tribal knowledge” to work. In other words, the engineers who built the workflow have a thorough understanding of it. It might even be a point of pride that they know the system so well that new hires routinely come to them with questions about it. The problem arises, as we note in “Interpreting a Data-Driven World,” when those gurus retire or move on to other positions inside or outside the company, taking that knowledge with them. That problem can be solved by documenting the process and updating it each time a new feature is added to the workflow, but that isn’t a priority for many engineers. It becomes another data point on the “to-do” list that gets ignored.
Another issue that arises is mission creep. Over time, different people add different data to the workflow, or create workarounds to make up for what previous employees might have once known how to do. Before too long, a simple homegrown system to make sure current files aren’t overwritten or common CAD files can be reused becomes a sprawling behemoth.
The Platform Approach
Software to help engineers manage the high-priority data they need right away and the long-term data they might need to find in the future has come a long way. As we explain in “Vendors Push Parts Reuse to the Next Level,” everything from built-in CAD and CAE features to stand-alone PDM and full PLM implementations now address common data management headaches such as finding and reusing parts or analyses.
Another historical issue to data management was a one-size-fits-all approach by software providers. PLM software vendors are building more flexibility into their products now. Templated approaches to data management that address specific industries can be further customized based on the data that people in different roles in specific companies want to see.
Still, data management needs run a broad gamut, from simple file permission controls to full PLM. There are a lot of options in between. It’s also a constantly changing target, which means engineers should think of data management as a process to implement, not a solution to buy.
“[The] bottom line is that PLM is hard and management must be convinced of its value so they will pay for it,” writes Technology Consultant Jay Swindle, in response to last month’s DE article “A Dead End for PLM?” “Everybody else had better get their best people together to figure out how these shrink-wrap box software product capabilities can best be mapped into their solution space in a way that justifies the implementation and product use cost. Good luck with that.”
It may be tempting not to think of data management at all. You could ignore what could be significant long-term benefits of mining and sharing data, and just focus on the data you need to make a product that meets requirements. But as Swindle concludes: “Success is possible and worthwhile. So to them that’s tryin’, hang in there!” I encourage you to follow that advice, and for them that’s not trying yet: Take the first steps. After all, yesterday’s ignored data becomes tomorrow’s high-priority problem. That and the smell of the remains of last night’s fish fry remind me: I need to take out the garbage. That counts as exercise, right?